> ## Documentation Index > Fetch the complete documentation index at: https://docs.getcollate.io/llms.txt > Use this file to discover all available pages before exploring further. # AI Studio > Build and manage AI agents, workers, personas, plugins, and skills in Collate AI Studio. # AI Studio AI Studio is your control center for building and managing two kinds of AI entities in Collate: **agents** and **workers**. * An **agent** performs metadata tasks independently: it reads your data assets, applies logic, and takes action (for example, writing descriptions or flagging PII). You run agents on a schedule through AI Automations, or call them directly through the AI SDK. * A **worker** is part of Collate AI Chat's agentic fleet. When you ask a question in chat, Collate routes it to the right worker automatically. Workers are chat-only; you can't call them from AI Automations. Both are built from the same underlying pieces: * **Plugins** connect it to the data sources and external tools it needs to do its job. * **Skills** control what specific actions it's allowed to take. * Agents use a **persona** to define how they think and communicate; workers use **Planner Routing** instead, to decide how to decompose and route each request. See [Agents vs Workers](/ai-2-0/collate-ai/ai-studio/agents#agents-vs-workers) for a full comparison. AI Studio You set all of this up in AI Studio. Start by choosing what you want to configure: Create and manage AI agents that automate metadata tasks such as documentation, quality, tiering, and PII classification. Configure the AI entities that power Collate AI Chat's agentic fleet. Define the role, expertise, and communication style an agent uses when performing tasks. Connect agents to external tools and data sources to extend what they can access and act on. Attach reusable capabilities to agents to control what actions they can perform.